Natural Language Inference On Qnli

评估指标

Accuracy

评测结果

各个模型在此基准测试上的表现结果

模型名称	Accuracy	Paper Title	Repository
FNet-Large	85%	FNet: Mixing Tokens with Fourier Transforms
RealFormer	91.89%	RealFormer: Transformer Likes Residual Attention
Nyströmformer	88.7%	Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention
Q-BERT (Shen et al., 2020)	93.0	Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT	-
DeBERTaV3large	96%	DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing
Charformer-Tall	91.0%	Charformer: Fast Character Transformers via Gradient-based Subword Tokenization
ELECTRA	95.4%	-	-
SMART-BERT	-	SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization
data2vec	91.1%	data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language
SpanBERT	94.3%	SpanBERT: Improving Pre-training by Representing and Predicting Spans
24hBERT	90.6	How to Train BERT with an Academic Budget
ASA + RoBERTa	93.6%	Adversarial Self-Attention for Language Understanding
PSQ (Chen et al., 2020)	94.5	A Statistical Framework for Low-bitwidth Training of Deep Neural Networks
TRANS-BLSTM	94.08%	TRANS-BLSTM: Transformer with Bidirectional LSTM for Language Understanding	-
T5-Small	90.3%	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
ASA + BERT-base	91.4%	Adversarial Self-Attention for Language Understanding
ALICE	99.2%	SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization
T5-Base	93.7%	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
T5-11B	96.7%	Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
ERNIE	91.3%	ERNIE: Enhanced Language Representation with Informative Entities

0 of 43 row(s) selected.